Commit graph

589,246 commits

Author SHA1 Message Date
Willem de Bruijn
2793a23aac net: validate variable length ll headers
Netdevice parameter hard_header_len is variously interpreted both as
an upper and lower bound on link layer header length. The field is
used as upper bound when reserving room at allocation, as lower bound
when validating user input in PF_PACKET.

Clarify the definition to be maximum header length. For validation
of untrusted headers, add an optional validate member to header_ops.

Allow bypassing of validation by passing CAP_SYS_RAWIO, for instance
for deliberate testing of corrupt input. In this case, pad trailing
bytes, as some device drivers expect completely initialized headers.

See also http://comments.gmane.org/gmane.linux.network/401064

Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-03-09 22:13:01 -05:00
Linus Torvalds
a6e434e955 Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Pull drm fixes from Dave Airlie:
 "A few imx fixes I missed from a couple of weeks ago, they still aren't
  that big and fix some regression and a fail to boot problem.

  Other than that, a couple of regression fixes for radeon/amdgpu, one
  regression fix for vmwgfx and one regression fix for tda998x"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  Revert "drm/radeon/pm: adjust display configuration after powerstate"
  drm/amdgpu/dp: add back special handling for NUTMEG
  drm/radeon/dp: add back special handling for NUTMEG
  drm/i2c: tda998x: Choose between atomic or non atomic dpms helper
  drm/vmwgfx: Add back ->detect() and ->fill_modes()
  drm/radeon: Fix error handling in radeon_flip_work_func.
  drm/amdgpu: Fix error handling in amdgpu_flip_work_func.
  drm/imx: Add missing DRM_FORMAT_RGB565 to ipu_plane_formats
  drm/imx: notify DRM core about CRTC vblank state
  gpu: ipu-v3: Reset IPU before activating IRQ
  gpu: ipu-v3: Do not bail out on missing optional port nodes
2016-03-09 19:12:37 -08:00
Linus Torvalds
8205ff1dc8 I previously sent a fix that prevents all trace events from being called
if the current cpu is offline. But I forgot that in 3.18, we added lockdep
 checks to test RCU usage even when the event is disabled. Although there
 cannot be any bug when a cpu is going offline, we now get false warnings
 triggered by the added checks of the event being disabled.
 
 I removed the check from the tracepoint code itself, and added it to the
 condition section (which is "1" for 'no condition'). This way the online
 cpu check will get checked in all the right locations.
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1
 
 iQEcBAABAgAGBQJW4N3rAAoJEKKk/i67LK/8WBIH/1WS6n919iU5sbdwmb843o4v
 KTeZ962l7fsiU+Op2ha4eLO6qXwa85X1sq3yHxo7APttE0oN933b6VQFjEs+HnqU
 INZorOQpy7soztHNewr48hdS0Z/x57xHywuf9i1K51zKCycuhupS6eZxN65zcuZp
 jeyg1dWqcvUzQcbxc5xflt0+n27txUpHix3e290aNoH9cya7gdbXi5dWAQgM8Kfm
 l8i2DJeyEy9nAKMjsKpKvdPkV5C8ZMGS1sJc/Psx9MGL08kM5Lqtuu8gkrvjqiLk
 HYmWPKQ+l3OROORd6Sia88SniPT9ZU4A73CobgPt5flr8BvU51kxLeObD8myXps=
 =s1tp
 -----END PGP SIGNATURE-----

Merge tag 'trace-fixes-v4.5-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace

Pull tracing fix from Steven Rostedt:
 "I previously sent a fix that prevents all trace events from being
  called if the current cpu is offline.

  But I forgot that in 3.18, we added lockdep checks to test RCU usage
  even when the event is disabled.  Although there cannot be any bug
  when a cpu is going offline, we now get false warnings triggered by
  the added checks of the event being disabled.

  I removed the check from the tracepoint code itself, and added it to
  the condition section (which is "1" for 'no condition').  This way the
  online cpu check will get checked in all the right locations"

* tag 'trace-fixes-v4.5-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
  tracing: Fix check for cpu online when event is disabled
2016-03-09 19:01:58 -08:00
Eryu Guan
6ffe77bad5 ext4: iterate over buffer heads correctly in move_extent_per_page()
In commit bcff24887d ("ext4: don't read blocks from disk after extents
being swapped") bh is not updated correctly in the for loop and wrong
data has been written to disk. generic/324 catches this on sub-page
block size ext4.

Fixes: bcff24887d ("ext4: don't read blocks from disk after extentsbeing swapped")
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2016-03-09 21:37:53 -05:00
Linus Walleij
cc998d8bc7 Linux 4.5-rc5
-----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1
 
 iQEcBAABAgAGBQJWyN0eAAoJEHm+PkMAQRiGqIAIAKKodaqX5ACJhTRozj3GN5iV
 dDHU/SJQj4nIyJecaCVAJIBa3gvAX6GyY+Jg4JKJ4TKAdR0Hd/3EwOWIR+0+BQIM
 0MqmB0CRLzq42AOQtpDUdwB+OTE8jFQFQd2gFKuQYJJ61ppykCC36OWV0bTfQLSV
 b2esO4Ry6eoQnDMw8oT52ncUIZEvQ2DZE3L6tNDEPD/0je14GWkV1Fx1+X2jb9cB
 diFA2TmaEEXMHNT1NCLSQ+D7QefXV3mFl85leNlFi5QQNy7ZdSh7kvvOodMQ2uAS
 qa9V8Uk6LZYv5O71+Jr5Rmlqh3GxNRCMXu2tlMd2gtw8ApEvBw6XoL5YZYE13Lk=
 =3HMg
 -----END PGP SIGNATURE-----

Merge tag 'v4.5-rc5' into devel

Linux 4.5-rc5
2016-03-10 09:29:25 +07:00
Linus Torvalds
380173ff56 Merge branch 'akpm' (patches from Andrew)
Merge fixes from Andrew Morton:
 "13 fixes"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
  dma-mapping: avoid oops when parameter cpu_addr is null
  mm/hugetlb: use EOPNOTSUPP in hugetlb sysctl handlers
  memremap: check pfn validity before passing to pfn_to_page()
  mm, thp: fix migration of PTE-mapped transparent huge pages
  dax: check return value of dax_radix_entry()
  ocfs2: fix return value from ocfs2_page_mkwrite()
  arm64: kasan: clear stale stack poison
  sched/kasan: remove stale KASAN poison after hotplug
  kasan: add functions to clear stack poison
  mm: fix mixed zone detection in devm_memremap_pages
  list: kill list_force_poison()
  mm: __delete_from_page_cache show Bad page if mapped
  mm/hugetlb: hugetlb_no_page: rate-limit warning message
2016-03-09 18:27:52 -08:00
Lars-Peter Clausen
7f8b8f3fba mpt3sas: Remove unnecessary synchronize_irq() before free_irq()
Calling synchronize_irq() right before free_irq() is quite useless. On
one hand the IRQ can easily fire again before free_irq() is entered, on
the other hand free_irq() itself calls synchronize_irq() internally (in
a race condition free way), before any state associated with the IRQ is
freed.

Patch was generated using the following semantic patch:
// <smpl>
@@
expression irq;
@@
-synchronize_irq(irq);
 free_irq(irq, ...);
// </smpl>

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-03-09 20:42:47 -05:00
Douglas Gilbert
5ecee0a3ee sg: fix dxferp in from_to case
One of the strange things that the original sg driver did was let the
user provide both a data-out buffer (it followed the sg_header+cdb)
_and_ specify a reply length greater than zero. What happened was that
the user data-out buffer was copied into some kernel buffers and then
the mid level was told a read type operation would take place with the
data from the device overwriting the same kernel buffers. The user would
then read those kernel buffers back into the user space.

From what I can tell, the above action was broken by commit fad7f01e61
("sg: set dxferp to NULL for READ with the older SG interface") in 2008
and syzkaller found that out recently.

Make sure that a user space pointer is passed through when data follows
the sg_header structure and command.  Fix the abnormal case when a
non-zero reply_len is also given.

Fixes: fad7f01e61
Cc: <stable@vger.kernel.org> #v2.6.28+
Signed-off-by: Douglas Gilbert <dgilbert@interlog.com>
Reviewed-by: Ewan Milne <emilne@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-03-09 20:41:04 -05:00
Dan Williams
489011652a Merge branch 'for-4.6/pfn' into libnvdimm-for-next 2016-03-09 17:15:43 -08:00
Zhen Lei
d6b7eaeb03 dma-mapping: avoid oops when parameter cpu_addr is null
To keep consistent with kfree, which tolerate ptr is NULL.  We do this
because sometimes we may use goto statement, so that success and failure
case can share parts of the code.  But unfortunately, dma_free_coherent
called with parameter cpu_addr is null will cause oops, such as showed
below:

  Unable to handle kernel paging request at virtual address ffffffc020d3b2b8
  pgd = ffffffc083a61000
  [ffffffc020d3b2b8] *pgd=0000000000000000, *pud=0000000000000000
  CPU: 4 PID: 1489 Comm: malloc_dma_1 Tainted: G           O    4.1.12 #1
  Hardware name: ARM64 (DT)
  PC is at __dma_free_coherent.isra.10+0x74/0xc8
  LR is at __dma_free+0x9c/0xb0
  Process malloc_dma_1 (pid: 1489, stack limit = 0xffffffc0837fc020)
  [...]
  Call trace:
    __dma_free_coherent.isra.10+0x74/0xc8
    __dma_free+0x9c/0xb0
    malloc_dma+0x104/0x158 [dma_alloc_coherent_mtmalloc]
    kthread+0xec/0xfc

Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Jan Stancek
86613628b3 mm/hugetlb: use EOPNOTSUPP in hugetlb sysctl handlers
Replace ENOTSUPP with EOPNOTSUPP.  If hugepages are not supported, this
value is propagated to userspace.  EOPNOTSUPP is part of uapi and is
widely supported by libc libraries.

It gives nicer message to user, rather than:

  # cat /proc/sys/vm/nr_hugepages
  cat: /proc/sys/vm/nr_hugepages: Unknown error 524

And also LTP's proc01 test was failing because this ret code (524)
was unexpected:

  proc01      1  TFAIL  :  proc01.c:396: read failed: /proc/sys/vm/nr_hugepages: errno=???(524): Unknown error 524
  proc01      2  TFAIL  :  proc01.c:396: read failed: /proc/sys/vm/nr_hugepages_mempolicy: errno=???(524): Unknown error 524
  proc01      3  TFAIL  :  proc01.c:396: read failed: /proc/sys/vm/nr_overcommit_hugepages: errno=???(524): Unknown error 524

Signed-off-by: Jan Stancek <jstancek@redhat.com>
Acked-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: David Rientjes <rientjes@google.com>
Acked-by: Hillf Danton <hillf.zj@alibaba-inc.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Ard Biesheuvel
ac343e882a memremap: check pfn validity before passing to pfn_to_page()
In memremap's helper function try_ram_remap(), we dereference a struct
page pointer that was derived from a PFN that is known to be covered by
a 'System RAM' iomem region, and is thus assumed to be a 'valid' PFN,
i.e., a PFN that has a struct page associated with it and is covered by
the kernel direct mapping.

However, the assumption that there is a 1:1 relation between the System
RAM iomem region and the kernel direct mapping is not universally valid
on all architectures, and on ARM and arm64, 'System RAM' may include
regions for which pfn_valid() returns false.

Generally speaking, both __va() and pfn_to_page() should only ever be
called on PFNs/physical addresses for which pfn_valid() returns true, so
add that check to try_ram_remap().

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Kirill A. Shutemov
0a2e280b6d mm, thp: fix migration of PTE-mapped transparent huge pages
We don't have native support of THP migration, so we have to split huge
page into small pages in order to migrate it to different node.  This
includes PTE-mapped huge pages.

I made mistake in refcounting patchset: we don't actually split
PTE-mapped huge page in queue_pages_pte_range(), if we step on head
page.

The result is that the head page is queued for migration, but none of
tail pages: putting head page on queue takes pin on the page and any
subsequent attempts of split_huge_pages() would fail and we skip queuing
tail pages.

unmap_and_move_huge_page() will eventually split the huge pages, but
only one of 512 pages would get migrated.

Let's fix the situation.

Fixes: 248db92da1 ("migrate_pages: try to split pages on queuing")
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Ross Zwisler
30f471fd88 dax: check return value of dax_radix_entry()
dax_pfn_mkwrite() previously wasn't checking the return value of the
call to dax_radix_entry(), which was a mistake.

Instead, capture this return value and return the appropriate VM_FAULT_
value.

Signed-off-by: Ross Zwisler <ross.zwisler@linux.intel.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Matthew Wilcox <willy@linux.intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Jan Kara
566e8dfd88 ocfs2: fix return value from ocfs2_page_mkwrite()
ocfs2_page_mkwrite() could mistakenly return error code instead of
mkwrite status value.  Fix it.

Signed-off-by: Jan Kara <jack@suse.cz>
Cc: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Joseph Qi <joseph.qi@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Mark Rutland
0d97e6d802 arm64: kasan: clear stale stack poison
Functions which the compiler has instrumented for KASAN place poison on
the stack shadow upon entry and remove this poison prior to returning.

In the case of cpuidle, CPUs exit the kernel a number of levels deep in
C code.  Any instrumented functions on this critical path will leave
portions of the stack shadow poisoned.

If CPUs lose context and return to the kernel via a cold path, we
restore a prior context saved in __cpu_suspend_enter are forgotten, and
we never remove the poison they placed in the stack shadow area by
functions calls between this and the actual exit of the kernel.

Thus, (depending on stackframe layout) subsequent calls to instrumented
functions may hit this stale poison, resulting in (spurious) KASAN
splats to the console.

To avoid this, clear any stale poison from the idle thread for a CPU
prior to bringing a CPU online.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Reviewed-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Mark Rutland
e1b77c9298 sched/kasan: remove stale KASAN poison after hotplug
Functions which the compiler has instrumented for KASAN place poison on
the stack shadow upon entry and remove this poision prior to returning.

In the case of CPU hotplug, CPUs exit the kernel a number of levels deep
in C code.  Any instrumented functions on this critical path will leave
portions of the stack shadow poisoned.

When a CPU is subsequently brought back into the kernel via a different
path, depending on stackframe, layout calls to instrumented functions
may hit this stale poison, resulting in (spurious) KASAN splats to the
console.

To avoid this, clear any stale poison from the idle thread for a CPU
prior to bringing a CPU online.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
Cc: Alexander Potapenko <glider@google.com>
Cc: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Mark Rutland
e3ae116339 kasan: add functions to clear stack poison
Functions which the compiler has instrumented for ASAN place poison on
the stack shadow upon entry and remove this poison prior to returning.

In some cases (e.g. hotplug and idle), CPUs may exit the kernel a
number of levels deep in C code.  If there are any instrumented
functions on this critical path, these will leave portions of the idle
thread stack shadow poisoned.

If a CPU returns to the kernel via a different path (e.g. a cold
entry), then depending on stack frame layout subsequent calls to
instrumented functions may use regions of the stack with stale poison,
resulting in (spurious) KASAN splats to the console.

Contemporary GCCs always add stack shadow poisoning when ASAN is
enabled, even when asked to not instrument a function [1], so we can't
simply annotate functions on the critical path to avoid poisoning.

Instead, this series explicitly removes any stale poison before it can
be hit.  In the common hotplug case we clear the entire stack shadow in
common code, before a CPU is brought online.

On architectures which perform a cold return as part of cpu idle may
retain an architecture-specific amount of stack contents.  To retain the
poison for this retained context, the arch code must call the core KASAN
code, passing a "watermark" stack pointer value beyond which shadow will
be cleared.  Architectures which don't perform a cold return as part of
idle do not need any additional code.

This patch (of 3):

Functions which the compiler has instrumented for KASAN place poison on
the stack shadow upon entry and remove this poision prior to returning.

In some cases (e.g.  hotplug and idle), CPUs may exit the kernel a number
of levels deep in C code.  If there are any instrumented functions on this
critical path, these will leave portions of the stack shadow poisoned.

If a CPU returns to the kernel via a different path (e.g.  a cold entry),
then depending on stack frame layout subsequent calls to instrumented
functions may use regions of the stack with stale poison, resulting in
(spurious) KASAN splats to the console.

To avoid this, we must clear stale poison from the stack prior to
instrumented functions being called.  This patch adds functions to the
KASAN core for removing poison from (portions of) a task's stack.  These
will be used by subsequent patches to avoid problems with hotplug and
idle.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Dan Williams
5f29a77cd9 mm: fix mixed zone detection in devm_memremap_pages
The check for whether we overlap "System RAM" needs to be done at
section granularity.  For example a system with the following mapping:

    100000000-37bffffff : System RAM
    37c000000-837ffffff : Persistent Memory

...is unable to use devm_memremap_pages() as it would result in two
zones colliding within a given section.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
Reviewed-by: Toshi Kani <toshi.kani@hpe.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Dan Williams
d77a117e68 list: kill list_force_poison()
Given we have uninitialized list_heads being passed to list_add() it
will always be the case that those uninitialized values randomly trigger
the poison value.  Especially since a list_add() operation will seed the
stack with the poison value for later stack allocations to trip over.

For example, see these two false positive reports:

  list_add attempted on force-poisoned entry
  WARNING: at lib/list_debug.c:34
  [..]
  NIP [c00000000043c390] __list_add+0xb0/0x150
  LR [c00000000043c38c] __list_add+0xac/0x150
  Call Trace:
    __list_add+0xac/0x150 (unreliable)
    __down+0x4c/0xf8
    down+0x68/0x70
    xfs_buf_lock+0x4c/0x150 [xfs]

  list_add attempted on force-poisoned entry(0000000000000500),
   new->next == d0000000059ecdb0, new->prev == 0000000000000500
  WARNING: at lib/list_debug.c:33
  [..]
  NIP [c00000000042db78] __list_add+0xa8/0x140
  LR [c00000000042db74] __list_add+0xa4/0x140
  Call Trace:
    __list_add+0xa4/0x140 (unreliable)
    rwsem_down_read_failed+0x6c/0x1a0
    down_read+0x58/0x60
    xfs_log_commit_cil+0x7c/0x600 [xfs]

Fixes: commit 5c2c2587b1 ("mm, dax, pmem: introduce {get|put}_dev_pagemap() for dax-gup")
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reported-by: Eryu Guan <eguan@redhat.com>
Tested-by: Eryu Guan <eguan@redhat.com>
Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Hugh Dickins
06b241f32c mm: __delete_from_page_cache show Bad page if mapped
Commit e1534ae950 ("mm: differentiate page_mapped() from
page_mapcount() for compound pages") changed the famous
BUG_ON(page_mapped(page)) in __delete_from_page_cache() to
VM_BUG_ON_PAGE(page_mapped(page)): which gives us more info when
CONFIG_DEBUG_VM=y, but nothing at all when not.

Although it has not usually been very helpul, being hit long after the
error in question, we do need to know if it actually happens on users'
systems; but reinstating a crash there is likely to be opposed :)

In the non-debug case, pr_alert("BUG: Bad page cache") plus dump_page(),
dump_stack(), add_taint() - I don't really believe LOCKDEP_NOW_UNRELIABLE,
but that seems to be the standard procedure now.  Move that, or the
VM_BUG_ON_PAGE(), up before the deletion from tree: so that the
unNULLified page->mapping gives a little more information.

If the inode is being evicted (rather than truncated), it won't have any
vmas left, so it's safe(ish) to assume that the raised mapcount is
erroneous, and we can discount it from page_count to avoid leaking the
page (I'm less worried by leaking the occasional 4kB, than losing a
potential 2MB page with each 4kB page leaked).

Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Geoffrey Thomas
910154d520 mm/hugetlb: hugetlb_no_page: rate-limit warning message
The warning message "killed due to inadequate hugepage pool" simply
indicates that SIGBUS was sent, not that the process was forcibly killed.
If the process has a signal handler installed does not fix the problem,
this message can rapidly spam the kernel log.

On my amd64 dev machine that does not have hugepages configured, I can
reproduce the repeated warnings easily by setting vm.nr_hugepages=2 (i.e.,
4 megabytes of huge pages) and running something that sets a signal
handler and forks, like

  #include <sys/mman.h>
  #include <signal.h>
  #include <stdlib.h>
  #include <unistd.h>

  sig_atomic_t counter = 10;
  void handler(int signal)
  {
      if (counter-- == 0)
         exit(0);
  }

  int main(void)
  {
      int status;
      char *addr = mmap(NULL, 4 * 1048576, PROT_READ | PROT_WRITE,
              MAP_PRIVATE | MAP_ANONYMOUS | MAP_HUGETLB, -1, 0);
      if (addr == MAP_FAILED) {perror("mmap"); return 1;}
      *addr = 'x';
      switch (fork()) {
         case -1:
            perror("fork"); return 1;
         case 0:
            signal(SIGBUS, handler);
            *addr = 'x';
            break;
         default:
            *addr = 'x';
            wait(&status);
            if (WIFSIGNALED(status)) {
               psignal(WTERMSIG(status), "child");
            }
            break;
      }
  }

Signed-off-by: Geoffrey Thomas <geofft@ldpreload.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Hillf Danton <hillf.zj@alibaba-inc.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-09 15:43:42 -08:00
Dan Williams
59e6473980 libnvdimm, pmem: clear poison on write
If a write is directed at a known bad block perform the following:

1/ write the data

2/ send a clear poison command

3/ invalidate the poison out of the cache hierarchy

Cc: <x86@kernel.org>
Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
Reviewed-by: Vishal Verma <vishal.l.verma@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2016-03-09 15:15:32 -08:00
Masanari Iida
0d6f3ebf9e Doc: nfs: Fix typos in Documentation/filesystems/nfs
This patch fix spelling typos found in Documentation/filesystems/nfs

Signed-off-by: Masanari Iida <standby24x7@gmail.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
2016-03-09 16:13:04 -07:00
Dan Williams
b5ebc8ec69 libnvdimm, pmem: fix kmap_atomic() leak in error path
When we enounter a bad block we need to kunmap_atomic() before
returning.

Cc: <stable@vger.kernel.org>
Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
Reviewed-by: Vishal Verma <vishal.l.verma@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2016-03-09 15:12:41 -08:00
NeilBrown
ff8e92d5d9 nvdimm/btt: don't allocate unused major device number
alloc_disk(0) does not require or use a ->major number,
all devices are allocated with a major of BLOCK_EXT_MAJOR.

So don't allocate btt_major.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2016-03-09 15:00:24 -08:00
NeilBrown
ec56151d38 nvdimm/blk: don't allocate unused major device number
When alloc_disk(0) is used ->major is completely ignored, all devices
are allocated with a "major" of BLOCK_EXT_MAJOR.

So don't allocate nd_blk_major

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2016-03-09 14:59:41 -08:00
Lv Zheng
c85cc817e5 ACPI / OSL: Add support to install tables via initrd
This patch adds support to install tables from initrd.

If a table in the initrd wasn't used by the override mechanism,
the table would be installed after initializing all RSDT/XSDT
tables.

Link: https://lkml.org/lkml/2014/2/28/368
Reported-by: Thomas Renninger <trenn@suse.de>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:55:02 +01:00
Lv Zheng
a543132ee0 ACPI / OSL: Clean up initrd table override code
This patch cleans up the initrd table override code by merging
redundant logics and re-ordering code blocks.

No functional changes.

Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:53:10 +01:00
Harb Abdulhamid
86e75410f0 PNP / ACPI: add ACPI_RESOURCE_TYPE_SERIAL_BUS as a valid type
An error message is printed for resources of type 19, which is a valid
supported resource type.  The Firmware Test Suite tool (fwts) reports
this as a test failure.  This change fixes the false test failures
for ASL that use type 19 (ACPI_RESOURCE_TYPE_SERIAL_BUS) resources.

Signed-off-by: Harb Abdulhamid <harba@codeaurora.org>
Signed-off-by: Timur Tabi <timur@codeaurora.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:50:55 +01:00
Colin Ian King
bea3c377c2 ACPI / util: remove redundant check if element is NULL
element is &package->package.elements[i] which can never be NULL
so the check to see if it is NULL is redundant and can be removed.

Detected with static analysis by CoverityScan

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:47:35 +01:00
Colin Ian King
b2ca5dae31 ACPI: Add acpi_force_32bit_fadt_addr option to force 32 bit FADT addresses
Some HP laptops seem to have invalid 64 bit FADT X_PM* addresses
which are causing various boot issues.  In these cases, it would
be useful to force ACPI to use the valid legacy 32 bit equivalent
PM addresses.  Add a acpi_force_32bit_fadt_addr to set the ACPICA
acpi_gbl_use32_bit_fadt_addresses to TRUE to force this override.

Link: https://bugs.launchpad.net/bugs/1529381
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:46:07 +01:00
Paul Gortmaker
75829dcf10 drivers/acpi: make pmic/intel_pmic_crc.c explicitly non-modular
The Kconfig currently controlling compilation of this code is:

drivers/acpi/Kconfig:config CRC_PMIC_OPREGION
drivers/acpi/Kconfig:   bool "ACPI operation region support for CrystalCove PMIC"

...meaning that it currently is not being built as a module by anyone.

Lets remove the couple modular references, so that when reading
the driver there is no doubt it is builtin-only.

Since module_init translates to device_initcall in the non-modular
case, the init ordering remains unchanged with this commit.

We also delete the MODULE_LICENSE tag etc. since all that information
is already contained at the top of the file in the comments.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Acked-by: Aaron Lu <aaron.lu@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:46:07 +01:00
Paul Gortmaker
020bf066a6 drivers/acpi: make apei/ghes.c more explicitly non-modular
The Kconfig currently controlling compilation of this code is:

config ACPI_APEI_GHES
      bool "APEI Generic Hardware Error Source"

...meaning that it currently is not being built as a module by anyone.

Lets remove the modular code that is essentially orphaned, so that
when reading the driver there is no doubt it is builtin-only.

Since module_init translates to device_initcall in the non-modular
case, the init ordering remains unchanged with this commit.

We replace module.h with moduleparam.h as we are keeping the
pre-existing module_param that the file has, as currently that is
the easiest way to maintain compatibility with the existing boot
arg use cases.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:46:07 +01:00
Paul Gortmaker
cc079f8cf7 drivers/acpi: make bgrt driver explicitly non-modular
The Kconfig for this driver is currently:

config ACPI_BGRT
    bool "Boottime Graphics Resource Table support"

...meaning that it currently is not being built as a module by anyone.
Lets remove all modular references, so that when reading the driver
there is no doubt it is builtin-only.

Since module_init translates to device_initcall in the non-modular
case, the init ordering remains unchanged with this commit.

We also delete the MODULE_LICENSE tag etc. since all that information
was (or is now) contained at the top of the file in the comments.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:46:07 +01:00
Prakash, Prashanth
f387e5b901 ACPI / CPPC: use MRTT/MPAR to decide if/when a req can be sent
The ACPI spec defines Minimum Request Turnaround Time(MRTT) and
Maximum Periodic Access Rate(MPAR) to prevent the OSPM from sending
too many requests than the platform can handle. For further details
on these parameters please refer to section 14.1.3 of ACPI 6.0 spec.

This patch includes MRTT/MPAR in deciding if or when a CPPC request
can be sent to the platform to make sure CPPC implementation is
compliant to the spec.

Signed-off-by: Prashanth Prakash <pprakash@codeaurora.org>
Acked-by: Ashwin Chaugule <ashwin.chaugule@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:35:29 +01:00
Prakash, Prashanth
beee23aebc ACPI / CPPC: replace writeX/readX to PCC with relaxed version
We do not have a strict read/write order requirement while accessing
PCC subspace. The only requirement is all access should be committed
before triggering the PCC doorbell to transfer the ownership of PCC
to the platform and this requirement is enforced by the PCC driver.

Profiling on a many core system shows improvement of about 1.8us on
average per freq change request(about 10% improvement on average).
Since these operations are executed while holding the pcc_lock,
reducing this time helps the CPPC implementation to scale much
better as the number of cores increases.

Signed-off-by: Prashanth Prakash <pprakash@codeaurora.org>
Acked-by: Ashwin Chaugule <ashwin.chaugule@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:35:29 +01:00
Prakash, Prashanth
8b0f578898 mailbox: pcc: optimized pcc_send_data
pcc_send_data() can be invoked during the execution of performance
critical code as in cppc_cpufreq driver. With acpi_* APIs, the
doorbell register accessed in pcc_send_data() if present in system
memory will be searched (in cached virt to phys addr mapping),
mapped, read/written and then unmapped. These operations take
significant amount of time.

This patch maps the performance critical doorbell register
during init and then reads/writes to it directly using the
mapped virtual address. This patch + similar changes to CPPC
acpi driver reduce the time per freq. transition from around
200us to about 20us for the CPPC cpufreq driver

Signed-off-by: Prashanth Prakash <pprakash@codeaurora.org>
Acked-by: Ashwin Chaugule <ashwin.chaugule@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:35:29 +01:00
Prakash, Prashanth
77e3d86f2f ACPI / CPPC: optimized cpc_read and cpc_write
cpc_read and cpc_write are used while holding the pcc_lock spin_lock,
so they need to be as fast as possible. acpi_os_read/write_memory
APIs linearly search through a list for cached mapping which is
quite expensive. Since the PCC subspace is already mapped into
virtual address space during initialization, we can just add the
offset and access the necessary CPPC registers.

This patch + similar changes to PCC driver reduce the time per freq.
transition from around 200us to about 20us for the CPPC cpufreq
driver.

Signed-off-by: Prashanth Prakash <pprakash@codeaurora.org>
Acked-by: Ashwin Chaugule <ashwin.chaugule@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:35:29 +01:00
Ashwin Chaugule
ad62e1e677 ACPI / CPPC: Optimize PCC Read Write operations
Previously the send_pcc_cmd() code checked if the
PCC operation had completed before returning from
the function. This check was performed regardless
of the PCC op type (i.e. Read/Write). Knowing
the type of cmd can be used to optimize the check
and avoid needless waiting. e.g. with Write ops,
the actual Writing is done before calling send_pcc_cmd().
And the subsequent Writes will check if the channel is
free at the entry of send_pcc_cmd() anyway.

However, for Read cmds, we need to wait for the cmd
completion bit to be flipped, since the actual Read
ops follow after returning from the send_pcc_cmd(). So,
only do the looping check at the end for Read ops.

Also, instead of using udelay() calls, use ktime as a
means to check for deadlines. The current deadline
in which the Remote should flip the cmd completion bit
is defined as N * Nominal latency. Where N is arbitrary
and large enough to work on slow emulators and Nominal
latency comes from the ACPI table (PCCT). This helps
in working around the CONFIG_HZ effects on udelay()
and also avoids needing different ACPI tables for Silicon
and Emulation platforms.

Signed-off-by: Ashwin Chaugule <ashwin.chaugule@linaro.org>
Signed-off-by: Prashanth Prakash <pprakash@codeaurora.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:35:28 +01:00
Zhiyi Sun
4db1fa6648 Documentation: kselftest: Remove duplicate word
Remove duplicate word "for" in kselftest.txt.

Signed-off-by: Zhiyi Sun <zhiyisun@msn.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
2016-03-09 15:33:38 -07:00
Javi Merino
c22d6bef7c doc: fix grammar
Some minor typos:

  - make is unbindable -> make it unbindable
  - a underlying -> an underlying
  - different version -> different versions

Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Javi Merino <javi.merino@arm.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
2016-03-09 15:33:06 -07:00
Philippe Loctaux
ce65508182 Documentation: Howto: Fixed subtitles style
Fixed subtitles style, aligned them with their header.

Signed-off-by: Philippe Loctaux <phil@philippeloctaux.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
2016-03-09 15:30:03 -07:00
Colin Ian King
0f9aeb7a6a ACPI / video: remove unused device_decode array
device_decode is now no longer used, so we may as well remove it.
Fixes gcc 6 warning:

drivers/acpi/acpi_video.c:221:19: warning: ‘device_decode’ defined
 but not used [-Wunused-const-variable]
 static const char device_decode[][30] = {
                   ^~~~~~~~~~~~~

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:29:07 +01:00
Oleg Drokin
1b6e75ee29 ACPI / EC: Deny write access unless requested by module param
In debugfs it's not enough to just set file mode to read-only to
deny write access to a file, instead just don't provide
the write method unless write access is really requested.

Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Acked-by: Thomas Renninger <trenn@suse.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:26:15 +01:00
Jean Delvare
87aca73737 NFC: microread: Drop platform data header file
Originally I only wanted to drop the unneeded inclusion of
<linux/i2c.h>, but then noticed that struct
microread_nfc_platform_data isn't actually used, and
MICROREAD_DRIVER_NAME is redefined in the only file where it is used,
so we can get rid of the header file and dead code altogether.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: Lauro Ramos Venancio <lauro.venancio@openbossa.org>
Cc: Aloisio Almeida Jr <aloisio.almeida@openbossa.org>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2016-03-09 23:25:45 +01:00
Kaiyen Chang
1d751584e0 ACPI / fan: Make struct dev_pm_ops const
Silence the following checkpatch warning:
WARNING: struct dev_pm_ops should normally be const.

Signed-off-by: Kaiyen Chang <kaiyen.chang@intel.com>
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2016-03-09 23:23:21 +01:00
Olof Johansson
1dea581f86 ARM: OMAP2+: critical DRA7xx fix for v4.5-rc
Force the DRA7xx Ethernet internal clock source to stay enabled
 per TI erratum i877:
 
 http://www.ti.com/lit/er/sprz429h/sprz429h.pdf
 
 Otherwise, if the Ethernet internal clock source is disabled, the
 chip will age prematurely, and the RGMII I/O timing will soon
 fail to meet the delay time and skew specifications for 1000Mbps
 Ethernet.
 
 This fix should go in as soon as possible.
 
 Basic build, boot, and PM test results are available here:
 
 http://www.pwsan.com/omap/testlogs/omap-critical-fixes-for-v4.5-rc/20160307014209/
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1
 
 iQIcBAABAgAGBQJW3UuoAAoJEMePsQ0LvSpL4JIP/j9A1ax1c6kGfNujSzBMrVL3
 I68l27ohfbo/MKMc/KsqkdahzGimIUmqkJGxrnA19UMhfyMb6l3pzlVswxfUUd10
 EXl/aKlPDa+Xl+A+TCwK78C69ZXHk4nqsNDSixuoIVfxM2uTZZZmNK3FOR+/EaQ8
 kUq3zwkg31HgsYxIyvqcCwpsxmDwKXx6fQ3sX5A6tQGvtsdeNofWJOVoGpZe0Ott
 tmt09VEvSGvXVEL1Um6U5A+8Mw6GPWa9/wih8nYaE70BswuOmIMUxeCkrShDadpn
 4Z8rqZg1Q8sdnI7ZCARS2WZ63+wrcjq04Yycf7m8feUc7cIDqlahWnrIWKuvpPAz
 P20LgrwRQDgifM2TzJupPRUKX+7BoACOXTIt2A3HuOIsZRfqysFx4qoOEdQNBlVq
 mOOwA/o8ly8hJI7uym8elrPY4MjZ4f6l2h/mFom0XrlS/1NcxXwuGqi9SJNneFSm
 ALyCIW7YnemoOex0tUcHUg2fiGaRceWlSmzHxI0WgVyOj86hrXc8OnpjnPmuhMQV
 i4pkL4Y1/UxZepd0b6QOTUC+LQvsWL008XLUr0SPm1d2Co9sxyzN8i0pXh07bsm0
 sOflS6DtwWSNenX/OVVQWk0r5amNwiFFpiw7tBWIeXYi4glhyizqdGjbzxRjxJUf
 QfFex23RAWtf/1ZrvqQO
 =kJw8
 -----END PGP SIGNATURE-----

Merge tag 'for-v4.5-rc/omap-critical-fixes-a' of git://git.kernel.org/pub/scm/linux/kernel/git/pjw/omap-pending into fixes

ARM: OMAP2+: critical DRA7xx fix for v4.5-rc

Force the DRA7xx Ethernet internal clock source to stay enabled
per TI erratum i877:

http://www.ti.com/lit/er/sprz429h/sprz429h.pdf

Otherwise, if the Ethernet internal clock source is disabled, the
chip will age prematurely, and the RGMII I/O timing will soon
fail to meet the delay time and skew specifications for 1000Mbps
Ethernet.

This fix should go in as soon as possible.

Basic build, boot, and PM test results are available here:

http://www.pwsan.com/omap/testlogs/omap-critical-fixes-for-v4.5-rc/20160307014209/

* tag 'for-v4.5-rc/omap-critical-fixes-a' of git://git.kernel.org/pub/scm/linux/kernel/git/pjw/omap-pending:
  ARM: dts: dra7: do not gate cpsw clock due to errata i877
  ARM: OMAP2+: hwmod: Introduce ti,no-idle dt property

Signed-off-by: Olof Johansson <olof@lixom.net>
2016-03-09 14:15:31 -08:00
David S. Miller
9531ab65f4 Merge branch 'kcm'
Tom Herbert says:

====================
kcm: Kernel Connection Multiplexor (KCM)

Kernel Connection Multiplexor (KCM) is a facility that provides a
message based interface over TCP for generic application protocols.
The motivation for this is based on the observation that although
TCP is byte stream transport protocol with no concept of message
boundaries, a common use case is to implement a framed application
layer protocol running over TCP. To date, most TCP stacks offer
byte stream API for applications, which places the burden of message
delineation, message I/O operation atomicity, and load balancing
in the application. With KCM an application can efficiently send
and receive application protocol messages over TCP using a
datagram interface.

In order to delineate message in a TCP stream for receive in KCM, the
kernel implements a message parser. For this we chose to employ BPF
which is applied to the TCP stream. BPF code parses application layer
messages and returns a message length. Nearly all binary application
protocols are parsable in this manner, so KCM should be applicable
across a wide range of applications. Other than message length
determination in receive, KCM does not require any other application
specific awareness. KCM does not implement any other application
protocol semantics-- these are are provided in userspace or could be
implemented in a kernel module layered above KCM.

KCM implements an NxM multiplexor in the kernel as diagrammed below:

+------------+   +------------+   +------------+   +------------+
| KCM socket |   | KCM socket |   | KCM socket |   | KCM socket |
+------------+   +------------+   +------------+   +------------+
      |                 |               |                |
      +-----------+     |               |     +----------+
                  |     |               |     |
               +----------------------------------+
               |           Multiplexor            |
               +----------------------------------+
                 |   |           |           |  |
       +---------+   |           |           |  ------------+
       |             |           |           |              |
+----------+  +----------+  +----------+  +----------+ +----------+
|  Psock   |  |  Psock   |  |  Psock   |  |  Psock   | |  Psock   |
+----------+  +----------+  +----------+  +----------+ +----------+
      |              |           |            |             |
+----------+  +----------+  +----------+  +----------+ +----------+
| TCP sock |  | TCP sock |  | TCP sock |  | TCP sock | | TCP sock |
+----------+  +----------+  +----------+  +----------+ +----------+

The KCM sockets provide the datagram interface to applications,
Psocks are the state for each attached TCP connection (i.e. where
message delineation is performed on receive).

A description of the APIs and design can be found in the included
Documentation/networking/kcm.txt.

In this patch set:

  - Add MSG_BATCH flag. This is used in sendmsg msg_hdr flags to
    indicate that more messages will be sent on the socket. The stack
    may batch messages up if it is beneficial for transmission.
  - In sendmmsg, set MSG_BATCH in all sub messages except for the last
    one.
  - In order to allow sendmmsg to contain multiple messages with
    SOCK_SEQPAKET we allow each msg_hdr in the sendmmsg to set MSG_EOR.
  - Add KCM module
    - This supports SOCK_DGRAM and SOCK_SEQPACKET.
  - KCM documentation

v2:
  - Added splice and page operations.
  - Assemble receive messages in place on TCP socket (don't have a
    separate assembly queue.
  - Based on above, enforce maxmimum receive message to be the size
    of the recceive socket buffer.
  - Support message assembly timeout. Use the timeout value in
    sk_rcvtimeo on the TCP socket.
  - Tested some with a couple of other production applications,
    see ~5% improvement in application latency.

Testing:

Dave Watson has integrated KCM into Thrift and we intend to put these
changes into open source. Example of this is in:

https://github.com/djwatson/fbthrift/commit/
dd7e0f9cf4e80912fdb90f6cd394db24e61a14cc

Some initial KCM Thrift benchmark numbers (comment from Dave)

Thrift by default ties a single connection to a single thread.  KCM is
instead able to load balance multiple connections across multiple epoll
loops easily.

A test sending ~5k bytes of data to a kcm thrift server, dropping the
bytes on recv:

QPS     Latency / std dev Latency
  without KCM
    70336     209/123
  with KCM
    70353     191/124

A test sending a small request, then doing work in the epoll thread,
before serving more requests:

QPS     Latency / std dev Latency
without KCM
    14282     559/602
with KCM
    23192     344/234

At the high end, there's definitely some additional kernel overhead:

Cranking the pipelining way up, with lots of small requests

QPS     Latency / std dev Latency
without KCM
   1863429     127/119
with KCM
   1337713     192/241

---

So for a "realistic" workload, KCM performs pretty well (second case).
Under extreme conditions of highest tps we still have some work to do.
In its nature a multiplexor will spread work between CPUs which is
logically good for load balancing but coan conflict with the goal
promoting affinity. Batching messages on both send and receive are
the means to recoup performance.

Future support:

 - Integration with TLS (TLS-in-kernel is a separate initiative).
 - Page operations/splice support
 - Unconnected KCM sockets. Will be able to attach sockets to different
   destinations, AF_KCM addresses with be used in sendmsg and recvmsg
   to indicate destination
 - Explore more utility in performing BPF inline with a TCP data stream
   (setting SO_MARK, rxhash for messages being sent received on
   KCM sockets).
 - Performance work
   - Diagnose performance issues under high message load

FAQ (Questions posted on LWN)

Q: Why do this in the kernel?

A: Because the kernel is good at scheduling threads and steering packets
   to threads. KCM fits well into this model since it allows the unit
   of work for scheduling and steering to be the application layer
   messages themselves. KCM should be thought of as generic application
   protocol acceleration. It to the philosophy that the kernel provides
   generic and extensible interfaces.

Q: How can adding code in the path yield better performance?

A: It is true that for just sending receiving a single message there
   would be some performance loss since the code path is longer (for
   instance comparing netperf to KCM). But for real production
   applications performance takes on many dynamics. Parallelism, context
   switching, affinity, granularity of locking, and load balancing are
   all relevant. The theory of KCM is that by an application-centric
   interface, the kernel can provide better support for these
   performance characteristics.

Q: Why not use an existing message-oriented protocol such as RUDP,
   DCCP, SCTP, RDS, and others?

A: Because that would entail using a completely new transport protocol.
   Deploying a new protocol at scale is either a huge undertaking or
   fundamentally infeasible. This is true in either the Internet and in
   the data center due in a large part to protocol ossification.
   Besides, KCM we want KCM to work existing, well deployed application
   protocols that we couldn't change even if we wanted to (e.g. http/2).

   KCM simply defines a new interface method, it does not redefine any
   aspect of the transport protocol nor application protocol, nor set
   any new requirements on these. Neither does KCM attempt to implement
   any application protocol logic other than message deliniation in the
   stream. These are fundamental requirement of KCM.

Q: How does this affect TCP?

A: It doesn't, not in the slightest. The use of KCM can be one-sided,
   KCM has no effect on the wire.

Q: Why force TCP into doing something it's not designed for?

A: TCP is defined as transport protocol and there is no standard that
   says the API into TCP must be stream based sockets, or for that
   matter sockets at all (or even that TCP needs to be implemented in a
   kernel). KCM is not inconsistent with the design of TCP just because
   to makes an message based interface over TCP, if it were then every
   application protocol sending messages over TCP would also be! :-)

Q: What about the problem of a connections with very slow rate of
   incoming data? As a result your application can get storms of very
   short reads. And it actually happens a lot with connection from
   mobile devices and it is a problem for servers handling a lot of
   connections.

A: The storm of short reads will occur regardless of whether KCM is used
   or not. KCM does have one advantage in this scenario though, it will
   only wake up the application when a full message has been received,
   not for each packet that makes up part of a bigger messages. If a
   bunch of small messages are received, the application can receive
   messages in batches using recvmmsg.

Q: Why not just use DPDK, or at least provide KCM like functionality in
   DPDK?

A: DPDK, or more generally OS bypass presumably with a TCP stack in
   userland, presents a different model of load balancing than that of
   KCM (and the kernel). KCM implements load balancing of messages
   across the threads of an application, whereas DPDK load balances
   based on queues which are more static and coarse-grained since
   multiple connections are bound to queues. DPDK works best when
   processing of packets is silo'ed in a thread on the CPU processing
   a queue, and packet processing (for both the stack and application)
   is fairly uniform. KCM works well for applications where the amount
   of work to process messages varies an application work is commonly
   delegated to worker threads often on different CPUs.

   The message based interface over TCP is something that could be
   provide by a DPDK or OS bypass library.

Q: I'm not quite seeing this for HTTP. Maybe for HTTP/2, I guess, or web
   sockets?

A: Yes. KCM is most appropriate for message based protocols over TCP
   where is easy to deduce the message length (e.g. a length field)
   and the protocol implements its own message ordering semantics.
   Fortunately this encompasses many modern protocols.

Q: How is memory limited and controlled?

A: In v2 all data for messages is now kept in socket buffers, either
   those for TCP or KCM, so socket buffer limits are applicable.
   This includes receive messages assembly which is now done ont teh
   TCP socket buffer instead of a separate queue-- this has the
   consequence that the TCP socket buffer limit provides an
   enforceable maxmimum message size.

   Additionally, a timeout may be set for messages assembly. The
   value used for this is taken from sk_rcvtimeo of the TCP socket.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2016-03-09 16:36:16 -05:00
Tom Herbert
10016594f4 kcm: Add description in Documentation
Add kcm.txt to desribe KCM and interfaces.

Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-03-09 16:36:16 -05:00